FI US Re Sip Rime Loan Loss Model Criteria 110815

Structured Finance
Residential Mortgage / U.S.A.

U.S. Prime RMBS Loan Loss Model Criteria

Sector-Specific Criteria
Fitch Ratings has finalized its new model framework for estimating losses on U.S. prime
mortgage pools collateralizing RMBS transactions. The core principle underpinning the
framework is the interaction between borrower equity and market value declines in determining
expected loss for each loan. In addition, the methodology considers loan-level attributes and
macroeconomic factors in deriving loss expectations.
Key Highlights
Forward-Looking and Countercyclical Framework: A key component of Fitchs new approach is
the application of its proprietary sustainable home price (SHP) model to adjust a propertys current
price to its sustainable value. The SHP component allows for a forward-looking, countercyclical view
on the potential for negative equity when projecting defaults and losses. With the SHP component,
collateral analysis, and economic considerations, the loss model produces higher credit
enhancement levels as risk enters the system and lower levels as risk neutralizes.
sLTV Leading Default Variable: The borrowers true equity in the property, as expressed by
the sustainable loan-to-value ratio (sLTV), is the most predictive default variable in the model.
The sLTV is calculated based on the lower of appraisal value (effectively the original CLTV)
and the value determined by Fitchs SHP model.
Intuitive Loss Severity Approach: Loss severity (LS) is calculated using an intuitive and
transparent accounting-based approach with core underlying assumptions calibrated to empirical
data. The LS calculation utilizes sustainable market value decline (sMVD) assumptions, quick sale
discounts, and foreclosure and liquidation timing and costs as inputs. Loan-level loss severities are
also subject to floors at each rating category to ensure minimum levels of credit enhancement and
provide for greater rating stability in stressed environments.
Transparent Rating Stresses: Fitch associates recent performance observations with an Asf
stress while AAAsf stresses are associated with the Great Depression. The model applies a
dynamic two-step process in determining sMVD stress assumptions whereby home prices are
first reduced to their sustainable value and then subjected to a further stressed sMVD
assumption that corresponds to each rating category. The AAAsf sMVD stress below
sustainable is associated with the most severe historical home price declines observed prior to
and during the Great Depression.
Models Role Limited: The RMBS loss model is only one component of Fitchs analysis. Other
critical factors of Fitchs RMBS analysis include the quality of the originator and servicer,
findings from third-party due diligence reviews, an analysis of the transaction cash flows, and a
review of the legal structure, including an assessment of the representations and warranties. It
is the combination of loan loss analysis using this new modeling framework and these other
risk elements that will inform a committee when assigning ratings to an RMBS transaction.
This report replaces Fitchs Feb. 1, 2011 exposure draft U.S. Prime RMBS Loan Loss Model
Criteria: Exposure Draft.

Inside
Key Changes Since Exposure Draft 2
Key Rating Drivers 2
Role of the Model in the Rating Process 3
Data Adequacy: Probability of Default 4
Probability of Default Analysis:
Drivers of Default 6
A Closer Look at Fitchs Sustainable
Home Price Model 8
Additional Probability of Default
Adjustments 13
Loss Severity 17
Generating Rating Stressed Losses 19
Treatment of Seasoned Loans 22
Sample Model Output 24
Sensitivity Analysis 26
Appendix A: Data Analysis and
Model Development 27
Appendix B: Roll Rate Analysis
Description 28
Appendix C: Regression Data and
Methodology 29
Appendix D: Economic Risk Factors 31
Appendix E: Heat Maps for Sustainable
Home Price Model (2006 and Current) 33
Appendix F: Summary of Changes to
Model Exposure Draft 34

Analysts
Rui Pereira
+1 212 908-0766
rui.pereira@fitchratings.com

Grant Bailey
+1 212 908-0544
grant.bailey@fitchratings.com
Sean Nelson
+1 212 908-0207
sean.nelson@fitchratings.com
Stefan Hilts
+1 212 908-9137
stefan.hilts@fitchratings.com
Suzanne Mistretta
+1 212 908-0639
suzanne.mistretta@fitchratings.com
Roelof Slump
+1 212 908-0705
roelof.slump@fitchratings.com
www.fitchratings.com August 15, 2011

U.S. Prime RMBS Loan Loss Model Criteria 2
August 15, 2011
Structured Finance
Key Changes Since Exposure Draft Publication
The core methodology has not materially changed since Fitch introduced the model in its
Feb. 1, 2011 criteria exposure draft; however, several key enhancements have been made to
better address some of the risk factors that drive Fitchs default and loss expectations. Chief
among the changes is the application of a two-step process to achieve a stressed sMVD
scenario, whereby home prices are first reduced to their sustainable value and then subjected
to a further stressed sMVD assumption at each rating category.
Fitch also combined two independent probability of default (PD) variables (sustainable market value
decline and original CLTV ratio) into a single new variable called sLTV, which is now the most
predictive variable of borrower default. Other changes made since the release of the criteria
exposure draft including the roll rate methodology used for determining expected defaults on
recent vintage collateral, the addition of separate loan-level PD adjustments for seasoned loans, and
changes to several default and loss variables are listed in Appendix F on page 34. Fitch will apply
the new model to analyze both new and existing ratings for prime RMBS transactions. The
application of the new model will negatively impact select existing prime RMBS ratings.
Key Rating Drivers
Fitchs updated U.S. RMBS loss model was developed to assess the credit risk of prime
residential mortgage collateral backing securitizations and covered bonds under base and
stressed home price and macroeconomic scenarios, at both the loan and pool levels.
Borrower home equity has been and will continue to be a primary driver of mortgage borrower
behavior. Home price projections are determined using a forward-looking, countercyclical
sustainable home price model. At the state level for each loan, the model calculates a sMVD,
which represents the difference between the home value at origination and what Fitch believes
to be the homes sustainable value. The sMVD is a significant driver in both the PD and LS
calculations. The major components of the model are summarized below.
Probability of Default
A regression-based analysis that estimates the probability of default based on 10 independent
variables Fitch found to strongly influence default risk. The variables include:
- A calculated sLTV assumption.
- Individual loan and borrower attributes.
- Regional economic risk factors.
In addition to the 10 independent variables that determine default risk, Fitch applies additional
PD penalty adjustments at the portfolio level to address concentration risks based on:
- The number of loans.
- The distribution of loan balances.
- The geographic composition of the pool.
PD penalties are also applied to loans with variability in repayment terms, such as hybrid ARMs
and interest-only mortgages. Fitch may also apply additional PD adjustments for
concentrations of borrower- or loan-related characteristics, such as multifamily properties, self-
employed borrowers, or first time homebuyers.
PD assumptions reflect an updated regression of mortgage performance data through April
2011. For purposes of the regression default data set, Fitch developed a roll-rate methodology
using five-year observed performance trends for estimating future defaults on loans still

Related Criteria
US RMBS Rating Criteria, Aug. 15,
2011
U.S. Residential Mortgage
Originator Review Criteria, June 30,
2011
U.S. Residential Mortgage Loan
Representations and Warranties
Criteria, June 30, 2011
U.S. Residential Mortgage Third-
Party Loan-Level Review Criteria,
June 30, 2011
U.S. RMBS Cash Flow Analysis
Criteria, July 8, 2011
U.S. Residential and Small Balance
Commercial Mortgage Servicer
Rating Criteria, Jan. 31, 2011
Criteria for Interest Rate Stresses in
Structured Finance Transactions,
March 21, 2011

August 15, 2011
Structured Finance
outstanding. As such, Fitchs updated regression not only considers cumulative defaults on
older vintages but also incorporates Fitchs cumulative default expectations for peak loss
vintage loans originated during the 20052007 period.
Loss Severity
LS is calculated using an accounting-based approach rather than the previous regression-
based model. The LS calculation utilizes sMVD assumptions, quick-sale discounts, and
foreclosure timing and cost assumptions as key inputs. Fitch believes this to be a more intuitive
and transparent approach, with all core underlying assumptions calibrated to empirical data.
Loan-level loss severities are also subject to floors at each rating category to ensure a
minimum amount of loss given default.
Each loans LS percentage represents the loss amount calculated for each loan (i.e. loan
balance less liquidation proceeds) expressed as a percentage of the loan balance.
Rating Stress Scenarios
The product of each loans PD and LS represents its base case loss expectation. Loss
expectations derived for each rating category above the base case are determined by applying
stresses to the calculated sMVD, liquidation timelines, and the economic risk factor.
Rating stresses applied to the propertys sustainable price are based on a two-step process.
Fitch first calculates a sMVD for each property using its SHP model and reduces its price to its
sustainable value, which is then reduced further by a fixed percentage that Fitch has
determined to be consistent with each rating category stress. For example, under the AAAsf
rating stress, Fitch assumes a propertys value will decline an additional 35% below its
sustainable value. While sMVD is the primary driver of the stress scenarios, additional stresses
are also applied to the economic risk factors and foreclosure timelines.
Seasoned Loans
When estimating losses on seasoned mortgage loans, Fitch uses two key adjustments the
borrowers updated sLTV and payment history to account for the additional observed
performance data. These adjustments can either increase or decrease the loans loss
expectation relative to loss estimates that would have otherwise been derived at origination.
Role of the Model in the Rating Process
Fitchs mortgage loss model allows the agency to express its credit opinion in a consistent manner
across pools with differing characteristics. While the model is an important component in the rating
process, Fitch factors in other important considerations when assigning ratings to a transaction.
These considerations include a review of the collateral performance history, the quality of loan
origination and servicing, results and findings of third-party due diligence reviews, an assessment of
the representations and warranties, legal structure, and deal documentation, and an analysis of the
cash flow structure. It is the combination of the pool-level loss analysis and these other elements
that will inform the committee when assigning ratings to a particular transaction.
While Fitch believes its updated framework will result in model output that is more consistent
with observed and projected defaults and losses on mortgage portfolios, the agency also
acknowledges it may not be applicable for select portfolios or transactions. In such cases, Fitch
will overlay additional considerations to address portfolio risk factors, utilize other analytical
Fitchs U.S. RMBS loss model and its
output is a single component in the
rating process, and Fitch considers
many factors before assigning ratings.

August 15, 2011
Structured Finance
approaches, apply rating caps in select cases as defined by criteria, or decline to rate the
transaction. For detailed criteria methodology on the individual components listed above, see
Fitchs Related Research on page 2 of this report.
Data Adequacy: Probability of Default
The PD component of the model estimates the probability a loan will default based on various
loan and borrower characteristics, as well as Fitchs sustainable home price forecasts and
macroeconomic factors. The model was developed using a logistic regression analysis on a
sample of 2.2 million prime, fixed-rate, fully amortizing loans originated between 1991 and
2007. For more information on the model development dataset and filters applied, see
Appendix A on page 27.
Fitch sought to incorporate as much empirical default data as possible, including the
performance of more recent vintages that have significantly underperformed during the housing
crisis. However, Fitch also acknowledges that prime collateral currently being originated under
more traditional underwriting standards is expected to record significantly stronger performance
compared with those recent vintages. In practical terms, this meant the model output needed to
produce expected loss output that appropriately reflects the heightened risk profile of the
20052007 vintage prime collateral but also sufficiently differentiates for newly originated
collateral underwritten to more traditional underwriting guidelines.
To this end, Fitch identified a portion of the development set with stronger collateral attributes that
the agency believes to be consistent with traditional prime loans (i.e. FICO > 720, full documentation,
CTLV < 80 if owner occupied, and CLTV < 70 if non-owner occupied). Roughly 25% of the original
2.2 million loan sample was deemed traditional, with the remaining 75% deemed nontraditional. To
lend more weight to the traditional loans, nontraditional loans were removed from the regression
development set such that the ratio of nontraditional to traditional loans went from 3:1 to 2:1.
Through the regression analysis, Fitch identified 10 significant drivers of default. The following
sections describe how default was defined and measured for purposes of the analysis and then
detail the 10 determinants of default.
Probability of Default Analysis: Measuring and Defining Default
To avoid placing too much emphasis on favorable historical default statistics, Fitchs dataset
consisted of loans originated from 1991 up to and including 2007. For the regression analysis,
Fitch considered any loan that had ever been in foreclosure (FC), was real estate owned
(REO), or had undergone a modification as a defaulted loan. Additionally, loans were
considered to have defaulted if their latest available delinquency status was 90 or more days
delinquent, regardless of whether paid off or still outstanding. This definition of default was
applied to all loans using data as of April 2011.
Because a considerable percentage of loans are still outstanding from the more recent vintages that
would not otherwise be flagged by the agencys default definition, Fitch utilized a roll-rate
methodology to project future defaults on loans originated between 2002 and 2007 that were
outstanding as of April 2011. Using historical default data alone would underestimate lifetime
defaults for these recent vintages and would bias the models default probability lower.
A roll-rate model was applied to recent
vintages to complete their default
picture. This allowed Fitch to use
cumulative default expectations for all
vintages as the dependent variable in
the regression.
Fitchs PD regression model
was developed using a dataset
of 2.2 million prime loans originated from
1991 through 2007, with performance
tracked through April 2011.

August 15, 2011
Structured Finance
Default projections for recent vintages were derived by analyzing the delinquency roll rate (the
rate at which loans move from one payment status to another, i.e. current to 30 days
delinquent) and default and prepayment performance of loans with similar characteristics over
the past five years. Fitch then used this analysis to extrapolate future defaults for outstanding
loans by assuming that performance continued at these historical rates for the remainder of the
life of the pool. Adjustments to future performance were made for projected changes in the
borrowers CLTVs based on assumptions of home price movements, amortization, and inflation.
For more details on Fitchs roll rate methodology, see Appendix B on page 28.
After projecting defaults on the outstanding collateral balance, those projected defaults were
added to actual defaults on non-outstanding loans to reach a cumulative default expectation for
each vintage. This allowed Fitch to use a cumulative default expectation for all vintages
1991 through 2007 as the dependent variable in its regression analysis. The charts on this
page and the top of page 6 show cumulative default expectations by vintage used by Fitch as
part of its regression analysis. These offer some perspective on the composition of Fitchs
cumulative defaults assumptions by vintage and highlight the importance of projected defaults
for more recent vintages. The chart on page 6 highlights the differential in cumulative default
expectations for the total sample versus the traditional loans subset.
Approximately one-third of Fitchs
expected defaults for the 20052007
peak vintages are projected to come
from loans that are now current.
Traditional prime loans with stronger
attributes have performed better in the
recent downturn and were considered in
the agencys cumulative default
expectations for this collateral subset.
0
3
6
9
12
15
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
0
20
40
60
80
100
Obser ved Def aul t s Pr oj ect ed Def aul t s Pool Fact or
Tradi t i onal Sampl e: Observed vs. Proj ect ed Def aul t s
Sour ce: Cor eLogi c/ Loan Per f or mance and Fi t ch Rat i ngs.
(%) (%)
6
7 8
0
5
10
15
20
25
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
0
20
40
60
80
100
Obser ved Def aul t s Pr oj ect ed Def aul t s %Pool Fact or (RHS)
Tot al Sampl e: Observed vs. Proj ect ed Def aul t s
Sour ce: Cor eLogi c/ Loan Per f or mance and Fi t ch Rat i ngs.
(%) (%)
14
21
23

Structured Finance

6
14
21
7 8
23
0
5
10
15
20
25
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
Tr adi t i onal Sampl e Tot al Sampl e
Tradi t i onal vs. Tot al Sampl e Act ual Plus Proj ect ed Def ault s
Sour ce: Cor eLogi c/ Loan Per f or mance.
(%)
Probability of Default Analysis: Drivers of Default
Fitch identified 10 key drivers, or variables, of default probability. Credit attributes of the
variables are used to determine their relative default risk. For categorical variables, one credit
attribute within each variable represents the baseline from which the relative risk of other
attributes is measured, holding all things constant. For example, the occupancy variable
consists of owner-occupied, second home, and investor property attributes. Owner occupied is
the baseline attribute from which the other two are measured in terms of higher (or with other
variables, sometimes lower) risk of default.
Fitch identified 10 key variables that
proved to be strong predictors of
borrower default.
Some variables are continuous such that the default risk rises or falls with changes in the
attributes. PDs derived for variables such as sLTV and credit scores reflect observed default
rates for each sLTV or score value. Higher default rates that were observed for the higher
sLTVs and lower credit scores are reflected by higher PDs for those loans. The following
section details each variable in order of importance for determining the base PD for each loan.
Probability of Default Risk Variables
Order of Influence Probability of Default Variables Loan Attributes
Baseline Attribute/
Probability of Default Relative to Baseline
1 sLTV
a
Continuous Higher sLTV = Higher PD
2 Credit Score (FICO) Continuous Higher Credit Score = Lower PD
3 Economic Risk Factor State- and MSA-Level Risk Multiplier Higher NRI = Higher PD
National Risk index Higher Multiplier = Higher PD
4 Loan Documentation Full to No Documentation Higher Fitch Score = Higher PD
5 Loan Purpose Purchase Baseline
Rate/Term Refinance Higher
Cash-Out Refinance Higher
6 Property Value Ratio Continuous Lower Value = Higher PD
7 Loan Term Term < 360 Months Lower
Term > 360 Months Higher
8 Back-End Debt-to-Income Ratio Continuous Higher DTI = Higher PD
9 Property Type Single Family/PUD Baseline
Condo Lower
Two to Four Family Higher
Co-Op Lower
10 Occupancy Owner-Occupied Primary Baseline
Second Home Higher
Investor (Non-Owner-Occupied) Higher
a
sLTV Sustainable loan-to-value ratio, which is original combined loan-to-value ratio adjusted by sustainable market value decline(sMVD). FICO Fair Isaac Corp.
PUD Planned unit development. DTI Debt-to-income ratio.

August 15, 2011

August 15, 2011
Structured Finance
Probability of Default Variable 1: Sustainable Loan-to-Value Ratio
When determining a loans default probability, the most predictive variable is borrower equity
throughout the life of the loan. Fitch considers borrower equity through its sLTV metric, which
measures the borrowers equity in the home calculated based on the lower of the purchase
price or appraisal value and the value determined by Fitchs SHP model. The SHP model
calculates the declines necessary to return to sustainable home prices (sMVD) at the state
level based on regional economic conditions and an analysis of fundamental price drivers. For
homes in markets considered to be overvalued by the SHP model, this means the PD model
will view the borrower as having less equity than the loan underwriting and original CLTV would
imply. Fitch assumes that the market value declines occur instantaneously with no timing
vectors employed.
Examining default performance across the development dataset, Fitch established there was a
strong correlation between observed home price declines and mortgage defaults. More
important than credit score, product type, documentation, or other borrower attributes, the
combination of original CLTV and sustainable market value declines (sMVD) explained a
majority of a loans behavior and is the strongest factor in determining a loans risk of default.
Fitchs analysis also showed this relationship held true across geographic regions, including
states with and without lender recourse against defaulting borrowers.
The sLTV variable reflects not only initial borrower equity but also a loans risk of default due to
market value adjustments needed to achieve price sustainability. The following chart illustrates
the relationship between sLTV and PD in the loan level dataset used for model estimation,
which includes projected defaults for the 2002 and later vintages. The bulk of observations with
greater than 100% sLTVs reflect 20042007 vintage loans, where prices were at their highest
at origination. The impact of subsequent declines is reflected in the high PDs for those buckets.
To account for nonlinearities in the relationship between sLTV and loan defaults, a logistic
transformation is applied to the sLTV variable in the default regression.
0
10
20
30
40
50
50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 135 140 145 150 155 160 165 170 175
sLTV as Proj ect ed at Ori gi nat i on
C
u
m
u
l
a
t
i
v
e

D
e
f
a
u
l
t
s
Incl udes Fi t ch's r ol l r at e pr oj ect i ons. sLTV Sust ai nabl e l oan-t o- val ue r at i o.
Sour ce: Cor eLogi c/ LoanPer f or mance and Fi t ch Rat i ngs.
Cumul at i ve Def aul t s by sLTV
( %)
a
a
The sustainable loan-to-value ratio
(sLTV) is the leading variable in Fitchs
new regression model. The sLTV
considers both the original CLTV as well
as state-level home price projections
from Fitchs SHP model.

August 15, 2011
Structured Finance

A Closer Look at Fitchs Sustainable Home Price Model
The SHP model is a regression-based model that aims to associate movements in home prices
with drivers fundamental to the housing market. An analysis of these drivers is the basis for
identifying deviations of prices from historic trends. Thus, the SHP model can identify housing
bubbles in the default and loss risk assessment for each loan. The key drivers used in the SHP
model are as follows and are listed roughly in order of importance:
- Unemployment Levels: Home prices are inversely related to unemployment levels,
meaning that as unemployment levels increase, home prices decline.
- Population and Housing Units: Home prices fluctuate with changes in the underlying
market dynamics; rising population is positively correlated with home prices, while increases
in housing supply negatively impact or slow home price growth.
- Household Income: As consumer purchasing power grows, home values will follow pace.
This relationship exists for both real and nominal income changes. For application in the
SHP model, home prices and incomes are incorporated as inflation-adjusted (real) indices.
- Mortgage Rates: Representing the cost of capital, mortgage rates are also inversely related
to home prices. A rise in mortgage rates constrains borrower affordability and decreases
long-term return on a property.
These fundamental drivers are used as inputs to the SHP model and are associated with a
coefficient (or multiplier) for determining the expected movement in home prices for any given
change in a driver. To reflect long-term trends, each factor is observed as a 16-quarter average,
including three years of historical and one year of projections. This helps ensure that a change in
sustainable prices predicted by the model truly reflects a change in long-term economics rather
than a short-term or seasonal movement. This regression is run at the state level to capture
differences that may arise in localized drivers. For example, unemployment is a weaker driver in
Florida, where a higher percentage than the national average of homeowners consists of fixed-
income retirees. However, floors are applied to unemployment, income, and population per unit
factor weights to ensure a minimum contribution from each of these factors, regardless of
regression fit. Additionally, the weight on mortgage rates is capped to prevent overfitting to this
component, which would cause sustainable values to be inflated in the current rate environment.
There is an embedded assumption in the SHP model that, over some period in the past 30 years, prices
have been sustainable within each state. This period varies from state to state but is typically associated
with regional or industry similarities. The model is fit only over periods where prices are considered to
have been sustainable. For example, California and much of the Northeast is fit over the period from
19792000, excluding 19881992, where a regional bubble and bust cycle was amongst the largest
observed before the recent downturn. Other fits include from 19872007 for the plains states and energy
producing states and 19872000 for states exposed to bubbles in both the mid 1980s and the 2000s.
The default state fit used for a plurality of states is without exclusions from 19792000.
By and large, home price movements in the 1980s and 1990s were highly correlated with this measure
of sustainable prices, with the notable regional exceptions described above. On average, prices
remained at sustainable levels until the 2000s, when market enthusiasm drove prices significantly higher,
resulting in aggressive overbuilding, which, in turn, caused a decline in the demand value of housing
stock. By 2006, the U.S. in the aggregate was approximately 42% overvalued in real terms by this
metric, with standout states such as California, Florida, and Arizona overvalued by more than 50% in real
terms. By 2010, home prices had dropped in California, Florida, and Arizona by 43%, 46%, and 48%,
respectively.
Fitch built a sustainable home price
model that associates movements in
home prices with fundamental drivers
like unemployment and supply and
demand dynamics.

August 15, 2011
Structured Finance
In practice, the SHP model is applied to each loan at origination, adjusting property appraisal
values to projected sustainable amounts, using the econometric data available at the time of
origination. As shown in the table to
the right for a 2006 vintage loan with
an original CLTV of 80% originated in
California, the SHP model implies a
sustainable value of one-half the
original value, which results in a 160%
Fitch-adjusted original CLTV or sLTV.
It is important to highlight that the
sLTV considers both the original CLTV
as well as the state-level home price
projection from Fitchs SHP model.
The sLTV is a key driver in the agencys default estimation.
Probability of Default Variable 2: Credit Score
Credit or FICO score remains a key driver of default in Fitchs model, as data continue to show
a strong relationship to default risk. As shown in the chart below, default risk is inversely
related to FICO score. All other variables remaining unchanged, a high borrower FICO score,
which indicates a sound repayment history of debt obligations, will result in a lower PD
A Closer Look at Fitchs Sustainable Home Price Model (continued)
Based on the SHP model, Fitch estimates that home prices remain overvalued by approximately 10%
nationally as of first-quarter 2011, with recent movements continuing to bring values closer to sustainable
levels. Regionally, significant differences remain, with plains states such as Texas and Nebraska
considered to be at sustainable levels, while much of the northeast remains more than 20% overvalued.
Californias level is similar to the national number, at a projected 11% decline. See Appendix E for more
information on state-level decline predictions.
The present version of the SHP model is able to project sustainable home prices at the state level,
providing the agency with a tool that can make differentiations across regional markets throughout the
country. However, the agency acknowledges the need for even greater granularity and expects to
expand its model to project sustainable home prices at the metropolitan statistical area (MSA) level in
the months ahead. In the interim, the agency may run sensitivities to evaluate the impact of pool
concentrations in MSAs that may differ from state-level sMVD assumptions. Commencing with the
publication of this criteria report, the agency will publish its updated sustainable home price views on a
quarterly basis.
sLTV Example

Property Value at Origination ($) 500,000
Loan Amount ($) 400,000
Original CLTV (%) 80
Projected Sustainable Value ($) 250,000
Sustainable Market Value Decline (%) 50
Sustainable Loan-to-Value Ratio (%) 160
sLTV Sustainable loan-to-value ratio. CLTV Combined loan-to-
value ratio.
0
50
100
150
200
250
300
350
1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009
H
o
u
s
i
n
g

I
n
d
e
x

V
a
l
u
e
Cal i f or ni a Sust ai nabl e Cal i f or ni a Case Shi l l er
Case Shi l l er Observed vs. Fi t ch Sust ai nabl e: Cal i f orni a
Sour ce: Case Shi l l er .
Pr i ces gener al l y at sust ai nabl e out si de
1990s and 2000s boom/ bust cycl es.

August 15, 2011
Structured Finance
assumption. Given the importance of FICO score in influencing default risk, Fitch will expect
new issue RMBS backed by seasoned pools to contain updated scores.
Probability of Default Variable 3: Economic Risk Factor
The economic risk factor (ERF) variable accounts for regional economic and demographic
factors as well as national macroeconomic trends. The regression analysis showed a strong
correlation between the ERF variable and default risk. In general, as the ERF increases,
indicating increased macroeconomic stress, default rates also increase. To account for
nonlinearities in the relationship between ERF and loan defaults, a logistic transformation was
applied to the ERF variable in the default regression. .
The volatility in macroeconomic indicators is most evidenced by the increase in the historical
NRI and ERF values beginning in mid-2004 as shown in the chart below. States that had
experienced very high default rates, such as California, exhibited very high ERF values prior to
and during the housing crisis. In contrast, Texas has exhibited relatively low ERF values and
mortgage defaults rates that have remained relatively stable throughout the downturn.
University Financial Associates, LLC (UFA), a mortgage portfolio analysis software provider
located in Ann Arbor, MI, provides Fitch with the quarterly ERF. The ERF reflects the impact of
economic factors and home price forecasts on future defaults and losses, which are
incorporated in UFAs National Risk Index (NRI) and regional risk (state and zip code)
multipliers. Assuming the same average loan credit quality, the NRI provides a default
probability for loans originated today relative to those during the 1990s. The NRI forecast
provided by UFA for the second quarter of 2011 was 1.41, indicating that newly originated
The Economic Risk Factor aligns Fitchs
default expectations with national
macroeconomic trends as well as
regional economic conditions.
0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
3. 0
3. 5
4. 0
4. 5
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
E
c
o
n
o
m
i
c

R
i
s
k

F
a
c
t
o
r
CA TX NRI NY
Sour ce: Uni ver si t y Fi nanci al Associ at es LLC
Hi st ori cal ERF and Nat i onal Ri sk Index
0
5
10
15
20
25
30
35
40
680 690 700 710 720 730 740 750 760 770 780 790 800 810 820
FICO Score
P
r
o
b
a
b
i
l
i
t
y

o
f

D
e
f
a
u
l
t
70%CLTV 80%CLTV
AAA Probabi l i t y of Def aul t Based on FICO Score and Ori gi nal Loan To Val ue
Not e: Assumes 45%sMVD, 3. 5 economi c r i sk f act or , f ul l document at i on, owner - occupi ed l oan.

August 15, 2011
Structured Finance
loans are expected to default at a rate 41% higher than the average of loans originated in the
1990s.
State- and zip code-level risk multipliers represent the level of expected risk over the life of a loan
relative to the national average on a constant quality basis. For example, if the UFA default multiplier
for a state is 0.90, expected defaults in that state are 90% of those for the average loan in the U.S.
Zip codes or states with a multiplier above 1.2 are deemed high risk of default. For more information
on the ERF methodology and performance in the regression, see Appendix D on page 30.
Probability of Default Variable 4: Loan Documentation
Fitchs analysis demonstrated that loans originated under reduced documentation programs have a
higher probability of default than loans underwritten to full documentation programs. Loans with no
verification of income or assets showed a very high propensity to default, particularly when
combined with other risk attributes such as lower FICO scores and high original CLTVs. Because
loan underwriting and origination practices can vary from one lender to another, Fitch utilizes a
scoring system based on four categories of documentation and verification standards for assessing
the risks associated with a lenders underwriting program and documentation practices.
The categories are differentiated by the type of verification and documentation of borrower
income, assets, employment, and reserves. In Fitchs opinion, these four elements are most
relevant for determining the quality of a lenders underwriting program and potential default risk.
Each category is assigned a weight to calculate a score on a five-point scale, with one
representing a fully documented loan and five representing a loan with very limited or no
documentation. Penalty factors applied to a loans base PD will be based on how that loan
scores on the scale, as shown in the table above.
While Fitch expects to see a very high percentage of full documentation loans in the current
environment, the agency also acknowledges that lenders continue to employ streamline
documentation programs for select borrowers. For such programs, Fitch will solicit additional
information from lenders to adequately assess eligibility guidelines, risk controls, and
performance. This will allow the agency to more accurately code loans originated under such
programs under its documentation scoring matrix. The originator review assessment and third-
party due diligence results will also impact the agency's opinion of such programs and loan
quality.
Probability of Default Variable 5: Loan Purpose
Fitch observed a default rate among cash-out refinance loans 100% higher than the purchase loan
baseline and, therefore, applies a PD that is 100% higher than that used for purchase loans.
Borrowers extracting equity from their home are often involved in debt consolidation or may be
experiencing other financial or personal hardships. If the borrower reloads debt after consolidation,
Documentation Categories and PD Multipliers
Fitch Documentation
Score Documentation Type
Probability of Default
Multiplier
<=1.5 Full documentation. 1.00
2.0 Income and employment verified; assets partially verified. 1.25
3.0 Income partially verified; employment and assets stated. 1.75
4.0 Income not verified, but some mitigants present . 2.25
5.0 Income not verified; inadequate mitigants. 2.75

August 15, 2011
Structured Finance
the debt burden may increase to a prohibitively high level and cause the borrower to default.
Additionally, borrowers using equity to finance large, non-essential expenses are more prone to
default if cash is otherwise needed due to a life event or change in financial circumstances. Rate
and term refinance mortgages are applied a 30% higher default rate compared to purchase loans.
Both cash-out and rate and term refinancings will be assigned higher PD penalties based on
the observed default behavior relative to purchase loans. Fitch believes that appraisals
associated with refinance loans may be less reliable than the purchase price, as there is no
market clearing bid to support the value.
Probability of Default Variable 6: Property Value Ratio
Fitch compared the value of each mortgage property to the median value at time of origination and
found this variable to be highly predictive of borrower default. Loans associated with property values
significantly below the median exhibited higher default rates relative to those at or above the median
value. This makes intuitive sense, as larger properties are generally associated with higher income
borrowers who may be less sensitive to income shocks than lower income borrowers. Less
desirable, low value properties may also increase the default risk if the borrower has more difficulty
selling the home. The property value to median variable is incorporated into Fitchs regression
model as a continuous variable.
Probability of Default Variable 7: Loan Term
Loans with less than or equal to 15-year terms exhibited a 50% lower default rate than the 30-year
term baseline; therefore, loans with maturities less than 30 years are assigned lower PDs to reflect
the lower probability of default. The shorter maturity results in faster amortization and equity build-up,
which increases a borrowers incentive to repay the loan. In particular, borrowers of a 15-year
mortgage voluntarily assume the higher payment despite having a smaller payment option with the
30-year mortgage, reflecting a positive selection bias. In contrast, Fitch applies a PD penalty of
150% relative to the 30-year baseline to loans with a maturity in excess of 30 years to account for
weaker performance, slower amortization, and heightened adverse selection risk.
Probability of Default Variable 8: Back-End Debt-to-Income Ratio
Default expectations are determined through analysis of back-end debt-to-income (DTI) ratios.
The back-end DTI is the ratio of the mortgage payment, taxes, and insurance plus other debt
obligations (i.e. auto loans and credit cards) to monthly gross income.
0
5
10
15
20
10-15 15-20 20-25 25-30 30-35 35-40 40-45 45-50 50- 55
DTI
%

L
o
a
n
s
0
5
10
15
20
P
r
o
b
a
b
i
l
i
t
y

o
f

D
e
f
a
u
l
t
%of Loans Pr obabi l i t y of Def aul t
Debt -t o-Income Rat i o and Probabi l i t y of Def aul t
Sour ce: Cor eLogi c/ LoanPer f or mance and Fi t ch Rat i ngs.
(%) (%)

August 15, 2011
Structured Finance
The chart above shows the positive relationship between DTI and default risk, where default
rates increase with higher DTIs. For DTIs above 55%, the sample consisted of less than 2% of
the DTI data set and exhibited slightly lower PDs than lower DTI buckets. Fitch attributes this
behavior to the small loan count of the data set and the possibility that borrowers with very high
DTIs consist of non-earner individuals primarily reliant on fixed income.
In Fitchs regression analysis, the DTI variable ranked eighth of the 10 variables in order of
importance for influencing default risk despite the very high default rates exhibited by
overleveraged borrowers concentrated in the 20062007 vintage collateral. Fitch attributes the
DTI variables lower ranking to the lack of quality data contained in the developmental dataset
(roughly 340,000 loans contained DTI information). However, as data quality improves and
consistent reporting of DTIs is put into practice, Fitch expects this variable to increase in
importance over time.
Probability of Default Variable 9: Property Type
The property type variable consists of single-family detached (SFD) homes, condominiums,
cooperatives, two- to four-family homes, and planned unit developments (PUDs). The SFD and
PUD property types are the baseline, since both have exhibited similarly low default rates historically.
Condominiums and co-ops exhibited lower default levels relative to the SFD/PUD baseline and
are respectively 10% and 30% less likely to default. Fitch believes these loans experience
fewer defaults because they are predominantly concentrated in heavily populated metropolitan
areas where demand is high. Also, in high-cost rental markets, condominiums provide
homeownership benefits with comparable monthly costs, which also fuels demand. In applying
the PD adjustments for these properties types, Fitch will consider the location of the properties.
In contrast, two- to four-family properties exhibited higher default rates compared with the baseline.
These homes are more prone to default risk since the borrower may be relying on income from
rental or other sources to help pay the mortgage. Likewise, the limited liquidity of these properties
also increases default risk. This property type is assigned a minimum penalty of 30%, which can be
increased to as high as four times the SFD baseline PD. The penalty will depend on the lenders
origination processes and underwriting guidelines, operating history, and portfolio performance.
Probability of Default Variable 10: Occupancy
Second homes are assigned a 10% higher default rate than the owner-occupied primary home
baseline. The higher PD reflects the increased likelihood of default on the second home if a
borrower is having financial difficulties and cannot sell it.
Investment properties exhibited a 35% higher likelihood of default than owner-occupied primary
home properties. Fitch attributes the high default rates to the effect of speculative investments and
the high-risk nature of rental properties. Speculative investing can increase default rates if the
property does not sell as fast as or at the price needed for an investor to break even. For rental
properties, the borrower is relying on income from external sources to repay the mortgage. The
investment property penalty may be increased above the minimum threshold noted above based on
the quality of the lenders underwriting guidelines and historical portfolio performance.
Additional Probability of Default Adjustments
In addition to the 10 variables in the regression model listed above, there are other important
factors that are significant determinants of default in Fitchs updated framework, including:

August 15, 2011
Structured Finance
- Products hybrid adjustable rate and interest-only mortgages.
- Pool concentration loan count and geographic concentration.
- Other qualitative considerations.
Separate loan-level PD assumptions (adjustments) are applied to adjustable rate mortgages
and other loan products that have variability in repayment terms to account for their higher
default risk. Fitch also makes adjustments at the pool level PD to account for loan count and
balance distribution risks and risks associated with geographically concentrated pools.
Hybrid Adjustable Rate Mortgages and Interest-Only Mortgages
Hybrid ARMs provide for a fixed rate of interest for a specified time, after which the interest rate
adjusts based on an index such as the six-month LIBOR or the Constant Maturity Treasury
(CMT). Principal is paid and amortizes beginning on the first payment date.
Interest-only (IO) mortgages can be either fixed or adjustable and provide for only interest to be
paid for a specified time, which typically ranges from five to 10 years. No principal is due to be
paid or amortized during the IO period. Once the IO period terminates, the payment converts to
a fully amortizing monthly amount. If the loan is both an ARM and an IO, the interest rate may
adjust prior to or concurrent with the end of the IO period. For fixed-rate IOs, the rate of interest
does not change when the IO period terminates; it is fixed at origination for the life of the loan.
The payment shock resulting from an increase in the monthly payment raises a borrowers
propensity to default. To address this risk, Fitch applies penalty multipliers to a loans base
case PD based on the size of the payment increase relative to the initial unadjusted payment
and the length of time before the first scheduled increase occurs.
To derive PD multipliers, Fitch analyzed the relative historical payment behavior of adjustable rate
and IO mortgage products that experienced an upward payment adjustment. This analysis was
used in lieu of a regression analysis due to the limited sample size of post-reset loans in a rising rate
environment. Prime hybrid ARMs and IOs originated in 2004 and later that have experienced a
payment adjustment have done so during an extremely low rate environment, and many have seen
a decline in their monthly payments, producing a distortedly low default profile for modeling
purposes. Furthermore, the higher level of defaults that has occurred before the adjustment date
indicates the presence of other risk factors as cause for default rather than payment shock alone.
Fitch accounts for payment reset risk in
ARM and IO products by applying PD
penalties that consider both the
magnitude of the payment shock, as
well as its timing.
PD Multipliers for ARMs and IOs
Projected Payment Increase (%) Time to Payment Increase (Years)
Average Probability of Default
Multiplier
<25 1 1.45
<25 3 1.30
<25 5 1.25
<25 7 1.20
<25 10 1.10
2550 1 1.85
2550 3 1.60
2550 5 1.45
2550 7 1.35
2550 10 1.20
>50 1 2.70
>50 3 2.20
>50 5 1.90
>50 7 1.65
>50 10 1.40

August 15, 2011
Structured Finance
For purposes of the analysis, Fitch established three bands of observed payment increases in
relation to the initial payment: less than 25%; 25%50%; and greater than 50%. Fitch then
analyzed the post-adjustment delinquency history of the ARM and IO dataset to determine the
PD multiplier based on the relative behavior of loans that had comparable credit characteristics
but faced different timing and amounts of payment adjustments.
The relative observed behavior of loans analyzed was used to determine PD penalty multipliers
for hybrid ARMs and IO mortgages. Fitch will estimate each loans future payment increase
based on mortgage loan terms and the Bsf interest rate stress detailed in its
March 11, 2011 criteria report Criteria for Interest Rate Stresses in Structured Finance
Transactions. The base case PD will be increased by the penalty that corresponds to the
payment increase derived from this approach.
The penalty multiplier determined by the projected payment shock is adjusted based on the
amount of time until the payment increase is expected to occur. For loans with a longer period
until the payment adjustment occurs, the penalty is reduced to reflect the greater likelihood that
the borrower will voluntarily prepay prior to experiencing an increase. Fitch will run multiple
prepayment scenarios to test the sensitivity of each loan and pool credit enhancement to the
prepayment assumption. Fitch may make adjustments to penalties to reflect individual lender
underwriting practices and performance history.
Pool Concentration Risk: Loan Count and Geographic
For the model output to reflect portfolio credit risk, there must be some consistency between the
RMBS transaction under analysis and the diversified portfolio of assets on which the model was
regressed. As an RMBS transaction becomes more concentrated, there is a greater risk the pool will
exhibit default characteristics that deviate from the default behavior of the total data set.
The updated U.S. RMBS rating framework seeks to identify and make adjustments for portfolio
concentrations that may impact performance. It does so through a combination of base
assumptions and an overlay of additional stresses, which act to increase credit enhancement if
a portfolio exhibits significant loan and/or geographic concentrations.
Concentration Risk 1: Loan Count
RMBS transactions with a small number of loans or those dominated by loans with very large
balances carry the risk that portfolio performance may be adversely impacted by a few assets
that may underperform relative to the statistically derived assumptions underlying their ratings.
This is simply because the variability of defaults inherently increases when a portfolio depends
on a smaller number of assets. In contrast, large diversified pools with more evenly distributed
balances benefit from the increased conformity to default estimates.
Fitchs model measures both loan and
geographic concentrations at the
portfolio level and applies penalties as
these risks increase.
Mortgage pools with weighted-average
loan counts below 300 will be subject to
PD penalties.
1. 0
1. 1
1. 2
1. 3
1. 4
150 200 225 275 300
WAN
P
o
o
l
-
L
e
v
e
l

P
D

A
d
j
u
s
t
m
e
n
t
Indi cat i ve AAA Loan Count Probabi l i t y of Def aul t Adj ust ment s
WAN Wei ght ed aver age number of l oans.

August 15, 2011
Structured Finance
Fitch uses the Herfindahl-Hirschman Index (HHI) to measure loan concentration in U.S. RMBS
transactions. The ratio is calculated as the sum of the squared pool shares (measured in fractions)
of each loan in the pool and expressed in a scale of zero to one, with values approaching zero
reflecting greater granularity. Fitch then uses the HHI index to calculate the weighted average
number (WAN) of loans (WAN=1/HHI) in the portfolio. The WAN accounts for both the number of
loans in the pool as well as the distribution of loan balances.
Pool granularity is established when each loans potential loss exposure represents a very small
share of the overall portfolio risk. Fitchs updated framework treats a pool with 300 evenly distributed
loans as being sufficiently granular and therefore does not apply an additional PD adjustment.
However, RMBS pools with an initial loan count or WAN below 300 loans will be subject to PD
penalties that are applied to the pools model-generated PD. The multiplier amount is inversely
related to the pools concentration metric and is stressed to a higher degree at higher rating
categories. For example, a high-quality RMBS pool with a 200 loan count as measured by its WAN
would be subject to a PD penalty of approximately 5% at the Bsf rating category and 20% at the
AAAsf rating category, compared with an identical pool that meets Fitchs minimum threshold.
RMBS pools where individual mortgage balances are disproportionately larger than the pools
average loan size are also subject to additional PD and LS sensitivity analysis. The results of
this additional analysis are applied outside the model and presented to the rating committee for
consideration before final ratings are assigned.
Concentration Risk 2: Geographic Distribution
Pools concentrated in a small number of geographic regions may be highly sensitive to
unforeseen localized stresses, including industrial shifts, deteriorating economic conditions, or
natural disaster events. While Fitch believes its SHP and ERF risk factors are adequate to protect
for a geographically diverse pool, some regions will naturally overperform or underperform
expectations. In a highly concentrated pool, performance is highly dependant on a small subset of
regions; therefore, loss protection derived by a diverse development dataset may prove
insufficient to cover the concentrated pools defaults and losses. To account for the additional risk
posed by highly concentrated pools, Fitch will apply PD multipliers to model results.
Fitch uses the HHI to measure the extent to which a pool has adequate spread of its
geographic risk. In this application, loans are identified as within one of the largest 50
MSA/CBSA regions in the U.S. or otherwise grouped by state, representing 100 total measured
regions. The HHI is calculated as the sum of the squared pool percentages within each region,
reflecting a weighted average pool concentration that ranges from 1% for a perfectly distributed
pool up to 100% for a pool concentrated in a single region.
Fitch applies a PD penalty multiplier to pools that are more geographically concentrated than the
aggregate development set. The magnitude of the penalty is determined by comparing the subject
pools HHI with the development sets HHI of 6.2%. The multiplier is applied as a sliding scale up to
a maximum increase of 5.0x the base PD for a pool with 100% concentration within a single region.
Fitch measures geographic
concentration risk and will apply
PD penalties to pools that are more
geographically concentrated than the
agencys aggregate development set.
PD Multipliers for Geographic Concentration
Percentile of Observed Prime Pools (%) Herfindahl-Hirschman Index (%) Probability of Default Multiplier
99.95 86.0 4.4
99.00 29.0 2.0
90.00 13.0 1.3
75.00 9.0 1.1
50.00 7.0 1.0
Maximum Penalty 100.0 5.0
19912007 Prime Pools

August 15, 2011
Structured Finance
As noted earlier in this report, model output must be considered along with qualitative
considerations. This is particularly important when looking at portfolios with significant
geographic concentrations. As such, in addition to applying the framework described above,
concentrations in select metropolitan statistical areas or states will result in additional scrutiny
on regional economies and housing markets and sensitivities for committee consideration.
Other Qualitative Considerations
Fitch conducted an analysis on a number of loan- and borrower-related variables that were not
included in the final PD regression. Factors such as risk premium, self-employed borrowers,
first-time homebuyers, and origination channel were considered for inclusion in the regression
but were ultimately excluded due to their statistical insignificance when controlled for other
attributes. However, to the extent that any of these attributes or characteristics is heavily
represented in a mortgage pool, Fitch will request additional underwriting and performance
information on these loans, which may result in adjustments to PD model output.
Loss Severity
Fitch revised its LS modeling framework by implementing an accounting-based approach that the
agency believes is more intuitive by taking into account the key underlying drivers of mortgage loss.
The agency calculates LS using regional projections of sMVD derived from its SHP model combined
with estimated carrying costs and liquidation expenses. A quick sale adjustment (QSA) discount is
applied to the sMVD-adjusted market value of the property to determine the recovery value based
on the observation that properties typically sell at a discount in distressed sales.
Foreclosure and liquidation costs, which include unpaid taxes, insurance, legal fees, and other
associated fees, are netted from recovery proceeds. Finally, carrying costs associated with
unpaid interest due on the loan also reduce total recoveries. After these adjustments, net
recoveries are subtracted from the loan balance to determine a loans loss severity, which is
expressed as a percentage of the loan balance.
While reverting to an accounting-based approach in estimating LS, Fitch considered empirical
data on actual observed severities through different economic scenarios in calibrating
Loss Severity Variables:
- Stressed sustainable market value
decline
- Quick sale adjustment
- Liquidation timelines and costs:
Fitch has transitioned from the previous
regression-based loss severity
approach to an accounting-based
framework that uses sMVD, foreclosure
and liquidation costs, and a quick sale
adjustment as key inputs:
Loss Severity Calculation Example
AAAsf Loss Severity Bsf Loss Severity
Appraisal/Sale Value ($) 600,000 600,000
Less: Stressed sMVD (35%)/(10%) 210,000 60,000
Less: QSA (15%) 58,500 81,000
Resale Value ($) 331,500 459,000

Resale Value Less Expenses
Liquidation Cost ($) 61,940 54,065
Carrying Cost ($) 72,000 36,000
Net Recovery ($) 197,560 368,935

Original Mortgage Amount ($) (OLTV = 80%) 480,000 480,000
Less: Net Recovery ($) 197,560 368,935
Loss Amount ($) 282,440 111,065

Loss Severity % (Loss Amount/Original Mortgage Amount) 59 23
Notes: Liquidation Costs: Legal costs of $5,000 and additional average cost of 1.7% of the appraised value per year
based on Fitchs assumed liquidation timelines (vary by state) to cover taxes and insurance, one-time repair cost of 1%
of resale property value, periodic maintenance of 0.25% per year based on Fitchs assumed liquidation timelines (vary by
state), and 6% of the resale value for sales commission. Carry Costs: AAAsf=36 months/Bsf=18 months of unpaid
interest on $480,000 at 5%.

August 15, 2011
Structured Finance
underlying assumptions. The framework recognizes the procyclical nature of defaults and
recoveries, with higher loss severities occurring during periods of higher defaults.
Loss Severity Variable 1: Sustainable Market Value Declines
In estimating the sale value of a property, appraised values are first adjusted by the sMVD,
which is estimated at the state level. The sMVD estimates are derived from Fitchs SHP model
and represent projected declines needed to achieve price sustainability as described in the
SHP model discussion on page 7. The sMVD assumptions applied in the LS calculations are
consistent with those used in Fitchs PD calculation. Fitch assumes that the market value
declines occur instantaneously with no timing vectors employed.
Loss Severity Variable 2: Quick Sale Adjustment
Fitch applies a QSA of 15% to all loans to reflect the discount on open-market values that sellers
of foreclosed properties may have to accept to quickly find interested buyers and to account for
distressed property conditions. The QSA discounts are applied as an incremental reduction to the
property price after it has been reduced by the sMVD and no timing vector is applied; Fitch
applies the QSA as an immediate reduction to the price.
Due to better average property conditions at liquidation, prime loans have lower observed QSAs
than Alt-A or subprime loans, which have historically experienced discounts closer to an average of
30%. QSA haircuts were tallied by comparing expected and observed losses for 90,000 liquidated
prime loans and calculating the price discount necessary to reconcile these numbers. Assumptions
for expected losses in this formula include carry costs, price declines, and all other foreclosure and
transaction costs, using observed information where available, and reflecting information from
Fitchs operation risk group and discussions with lenders and other market participants.
Loss Severity Variable 3: Liquidation Timelines and Costs
Fitch subtracts liquidation costs from the recovery amount expected to be realized at property
sale or disposition. Fitch estimates both fixed and variable liquidation costs based on available
industry information as well as information obtained from Fitchs operational risk group and
discussions with Fitch-rated servicers. The fixed cost assumptions reflect legal fees, while
variable costs are inclusive of real estate agent fees of 6% of the property sale price and
maintenance costs, which are assumed to be an initial 1% of the property value plus 0.25% per
year. Fitch also assumes taxes and insurance based on the propertys location and value, which
average approximately 2% annually and accrue monthly throughout the liquidation period.
To calculate carrying costs, Fitch applies the contractual interest rate of the loan for the duration of
the foreclosure period, which varies by state due to differences in local statutes and housing market
dynamics.
Servicer Advancing Considerations
Most U.S. RMBS transactions include a servicer advancing structure whereby the servicer will
advance borrower delinquent principal and interest to the trust to the extent that those amounts are
deemed recoverable. The servicer is generally able to recoup these advances from future mortgage
payments or the ultimate liquidation of the property. The advancing mechanism provides liquidity to
the transaction and allows for timely cash flows to investors. For transactions with servicer
advancing, Fitch believes that no additional liquidity is needed to pay interest on the bonds as long
as the financial condition of the servicer or backup servicer is supportive of the bonds highest rating

August 15, 2011
Structured Finance
and consistent with Fitchs counterparty criteria. To the extent that future RMBS transactions include
more restrictive or partial servicer advancing mechanisms, Fitch will evaluate their implications on
the liquidity of the transaction and make adjustments in its analysis accordingly.
Generating Rating Stressed Losses
A core objective of the model framework is to more clearly associate rating stress levels with
economic conditions and home price correction scenarios to aid investors in understanding the
type of stress and the associated loss expectations each rating category represents.
When deriving a loans PD and LS at each rating category, the primary variable used to
determine the stressed scenarios is the sMVD. Additional stresses are also applied to the ERF,
liquidation timelines, and loss severity floors as described below.
Rating Stress 1: Stressed sMVD
Fitchs sMVD rating stresses are designed to be countercyclical and consider positioning in the
housing market cycle. sMVD stresses will increase in unsustainable price environments and decline
as the housing market corrects and approaches sustainable levels. In application, Fitch applies a
two-step process whereby properties are first adjusted to their sustainable values through
application of the sMVD and then applied an
additional sMVD stress that corresponds to
different rating scenarios.
sMVD stresses influence both PD and LS
model calculations. When calculating a loans
PD, each loan is subjected to the sMVD rating
category stress, which increases its sLTV and,
therefore, its default probability. In estimating
LS, the stressed sMVDs are applied as
haircuts to the property value, thereby increasing the loan-level loss expected at each rating
category.
Step 1: The first step in Fitchs analysis is unwinding any perceived overvaluation in the regional
housing market by applying the SHP model adjustment described on page 7. By comparing actual
property prices to those indicated by the SHP model, Fitch derives a sMVD opinion in the base case
for each property at the state level.
Stressed Variables:
- Sustainable market value decline.
- Economic risk factor.
- Liquidation timelines.
- Loss severity floors.
Stressed sMVD by Rating
Category
Stresses Below Sustainable Level (%)
AAAsf 35
AAsf 30
Asf 25
BBBsf 20
BBsf 15
Bsf 10

Example of Sustainable Market Value Decline Stresses
(Assumes 15% Overvalued Property)
Property Value
Index
Base Case Bsf Asf AAAsf
120
102
92 10%
77
66
Total sMVD
Stresses
15% 22% 36% 45%
15% 15%
Current Value
Sustainable Value
15% 15%
25%
Rating Stress Scenarios
35%
Step 1,
Reduce property
value to
sustainable value
using SHP model.
Step 2,
Apply sMVD
rating stresses to
the sustainable
value.

August 15, 2011
Structured Finance
Step 2: After adjusting property values to sustainable levels, Fitch applies incremental stresses
to property prices that correspond to the different rating categories. The magnitude of each
rating categorys decline below sustainable levels is shown in the Stressed sMVD by Rating
Category table on the previous page, which was established by associating the AAAsf rating
category stress with the most severe declines observed in the early 1900s and during the
Great Depression.
For example, in the chart above, the SHP model would calculate a property in 2003 to be 20%
overvalued (base case sMVD = 20%). To determine that propertys AAAsf stressed sMVD, its
current price is first reduced by 20% (step 1) to achieve its sustainable price level. Then, the AAAsf
rating stress, an additional 35% decline to the sustainable price, is applied (step 2). This results in an
all-in 48% AAAsf stressed sMVD from the 2003 current value. In contrast, if a property is currently
at its sustainable value, step 1 is not necessary and the AAAsf stressed sMVD reflects only the
step 2 adjustment below sustainable (35%). This framework allows Fitch to make clear
differentiations in its sMVD stresses across regional housing markets and at different points in the
economic and housing cycle.
Rating Stress 2: Economic Risk Factor
The ERF stresses are applied only to the PD calculations and do not apply to loss severity.
Fitch has established ERF floors for
each rating category to provide more
stability in credit enhancement levels
and ratings during periods of
macroeconomic and unemployment
stress. The Asf ERF floor of 2.5 is
benchmarked against an economic
environment that includes high levels of
unemployment and roughly 25%
overvaluation in home prices to provide a combined rating stress that reflects a sufficiently
remote probability of occurrence.
Each loans base case ERF is compared with the ERF floor, and the higher level is used. The
stressed level is evaluated with each quarterly update from UFA. UFA periodically re-estimates its
Economic Risk Factor Floors
Rating Category Stress Minimum ERF Applied
AAAsf 3.5
AAsf 3.0
Asf 2.5
BBBsf 2.0
BBsf 1.5
Bsf 1.0

For macroeconomic and
unemployment stress scenarios, Fitch
benchmarked the economic scenario
experienced by the 2007 origination
vintage to an Asf. The ERF for the
2007 vintage was 2.4.
Example of sMVD Rating Stress Methodology
Through the Housing Cycle
35% 42% 54% 48%
50
60
70
80
90
100
110
120
130
140
150
2003 2006 2011 At Sustainable

H
o
m
e

P
r
i
c
e

I
n
d
e
x
Initial Price Sustainable 'AAAsf' Value Total 'AAAsf' sMVD Stress
Declines to
Sustainable
AAAsf
Below
Sustainable

August 15, 2011
Structured Finance
underlying models as additional data are received. Since the release of Fitchs exposure draft, UFA
has updated its model to reflect strategic defaults and estimates of borrower behavior among those
with significant negative equity. The ERF floors for the final model reflect the rescaling by UFA.

Rating Stress: How is Unemployment Stressed?
While unemployment is not stressed as an independent metric for determining loan performance, it is a core
component in both the sMVD and the ERF variables, which are both subject to stresses when generating
loss expectations for each rating category.

The correlation between unemployment and the sMVD assumptions can be illustrated with historical
economic downturns similar to Fitchs forward-looking investment-grade stress scenarios (which assume
home prices drop 20%35% below sustainable values). During the Great Depression, when home prices
declined 30%, the national unemployment rate rose to approximately 24% from 3%.

More recently, local regions that experienced market value declines below a sustainable level similar to
Fitchs high-investment grade stresses also experienced significant increases in unemployment levels. As
shown in the table below, the local unemployment rates increased to between 13% and 22%. Generally,
Fitch estimates that the unemployment rate would have to reach approximately 20% to result in a price
decline of 35% below a sustainable level.

Recently Observed Economic Experiences Similar to Fitch's Forward-Looking Investment-Grade Stress
Scenarios (%)
MSA
Home Price Decline Below
Fitch Sustainable Level
Unemployment Rate Pre-
Crisis
Unemployment Rate
Crisis Peak
Bakersfield-Delano, CA (41) 6 18
Merced, CA (37) 7 22
Lakeland-Winter Haven, FL (37) 3 13
Ocala, FL (23) 3 14

Rating Stress 3: Liquidation Timelines
Rating category stresses include assumptions made about liquidation timelines and costs,
which are based on various sources of information, including loan-level foreclosure data and
information gathered from Fitchs
operational risk group, agency
guidelines (Fannie Mae and Freddie
Mac), mortgage lenders, and other
sources. The recent downturn in the
U.S. housing market resulted in a
backlog of foreclosure proceedings
and distressed property sales that
have significantly increased resolution
timelines. In addition to rising defaults,
government initiatives (including loan modification schemes and more recent probes into lender
foreclosure processes that resulted in temporary moratoriums) have also contributed to the
extended foreclosure and liquidation timelines.
Fitch has accounted for these trends in its model and has benchmarked the Asf rating stress
scenario to the average timelines experienced in the current environment. All timeline
assumptions are state specific, reflecting differences in judicial and nonjudicial foreclosure
practices.
For markets with prices at sustainable
levels today, Fitch estimates that the
unemployment level would have to
reach 20% to result in a price decline of
35%.
Liquidation Timeline Stresses
Rating Category Average Timeline Assumed in Months
a

AAAsf 36
AAsf 30
Asf 24
BBBsf 22
BBsf 20
Bsf 18
a
Includes preforeclosure, foreclosure, and real-estate owned
timelines.

August 15, 2011
Structured Finance
Rating stress scenarios higher than Asf assume that timelines extend by up to a year from
todays timelines, while rating stress scenarios below Asf assume timelines improve by up to
six months. Fitch expects that newly originated loans will, on average, not enter delinquency for
several years from today, and will face shorter average timelines than those observed today
primarily due to a reduced supply of distressed inventory.
Rating Stress 4: Loss Severity Floors
Fitch also incorporated LS floor assumptions into its model that are scaled by rating category.
The floors range from 35% in the AAAsf stress to 15% in the B scenario, which is consistent
with minimum severities observed during periods of relatively low macroeconomic stress. This
mechanism results in minimum
thresholds of credit enhancement,
particularly for transactions with
concentrations of low CLTV loans.
Fitch believes that in investment-
grade stress scenarios, property
values could become considerably
more volatile; therefore, loss
protection should be sufficient to
protect against higher-than-expected
market value declines.
With respect to its application, the LS component of the model will produce two stressed levels at
each rating category. The first is generated by stressing the components of the loss severity
calculation (sMVD, foreclosure timelines, and the QSA), and the second is simply Fitchs defined
floor for that rating category. The model compares the two stressed LS values and uses the higher
of the two.
Treatment of Seasoned Loans
When estimating losses on seasoned mortgage loans, Fitch uses the same PD, LS, and rating
stress framework as that used for analyzing newly originated loans. However, two key
adjustments based on the borrowers updated sLTV and payment history are made to account
for the additional observed performance data available for seasoned loans. Depending on
Fitchs view of the borrowers equity position and the loans performance over time, these
adjustments can either increase or decrease the loans loss expectation compared to the loss
level the model would have assigned at loan origination.
In the process of developing its approach to analyzing seasoned loans, Fitch considered
several loan and borrower attributes it deemed predictive of default at origination but that could
potentially lose predictive ability as a loan seasons. Attributes such as loan documentation,
purpose, occupancy, and term were all explored for potential revision. The analysis showed
that all attributes retained their relevance to default behavior over time. For instance, the
performance of an average low-documentation loan compared to a full-documentation loan
was roughly the same at month 24 as it was at month 60. As such, Fitch uses all original
loan/borrower attributes when estimating losses for seasoned loans, except for sLTV.
Fitch uses a consistent model
framework for seasoned loans, with
adjustments for borrower equity
movements and loan performance.
Loss Severity Floors
Rating Category Minimum LS Assumed (%)
AAAsf 35
AAsf 30
Asf 25
BBBsf 20
BBsf 17
Bsf 15

August 15, 2011
Structured Finance
Updated sLTV
Fitch adjusts the sLTV of each seasoned loan to reflect the current equity position of the borrower.
The property value is adjusted to account for nominal home price movements and the current
loan balance is used to reflect any amortization since the time of loan origination. The values are
updated using the Case Shiller home price index, new appraisals, or other valuation methods
Fitch deems acceptable. Additionally, to the extent that Fitchs view on the propertys sustainable
value has fluctuated over time, the sLTV also will be adjusted accordingly.
Assuming the sustainable property value remains relatively stable in real terms, a borrowers
sLTV will generally decrease over time as inflation increases the actual nominal property value,
and the loan balance decreases with amortization. A lower sLTV translates into lower expected
losses, since sLTV is a prominent variable in PD. Loan seasoning combined with increased
borrower equity (through price inflation and loan amortization) will generally result in lower
expected losses.
Fitch observed this improvement in default rates over time when analyzing actual loan-level
performance data and considered several factors that may be driving the behavior. However,
after controlling for credit, economic environment, and, most importantly, borrower CLTV,
performance no longer improved as a loan seasoned; it remained relatively constant and, in
some cases, deteriorated over time.
As a result of this analysis, Fitch concluded that improvement in default rates over time was the
direct result of increased borrower equity. By simply updating the sLTV of a seasoned loan, the
model appropriately replicates this behavior and implicitly provides a so-called seasoning
benefit without having to dedicate a separate variable to do so.
Loan Delinquency and Payment History Adjustments
For loans that are past due or have recent delinquency, Fitch makes several adjustments to the
expected loss inputs to account for the increased risk that these loans present. The default
probability for delinquent (DQ) loans is determined by a regression-based model. The DQ
model was developed using three variables: the loans recent payment history; its most recent
delinquency status; and the base-case PD generated by the model, assuming the loan had a
clean payment history.
PD percentages progressively increase for higher rating category scenarios to reflect the
greater propensity of past due loans to be foreclosed under stressed conditions. Note that the
model not only increases the PD for loans that are currently delinquent; it also penalizes loans
that are performing but have some amount of recent delinquency.
In addition to PD adjustments, the model also treats DQ loans differently in terms of their LS
variables. For newly originated and performing seasoned loans, Fitch assumes base case
liquidation timelines lower than current averages, since the loans are generally not expected to
become delinquent for several years, when Fitch expects timelines will have improved from
todays levels. However, for DQ loans the timelines are extended to reflect timelines observed
in the current environment. Average timelines for DQ loans range from 24 months in the base
case to 36 months at the AAAsf rating stress.
The final adjustment made to DQ loans is to the stressed sMVD below the sustainable value.
This change arises from the fact that DQ loans have a shorter average time to liquidation than
newly originated or seasoned performing loans. Because of the shorter liquidation window,
Fitch believes DQ loans have a more limited range of property declines that are likely to occur
Updating the sLTV of a seasoned loan
results in an implicit seasoning benefit.
Fitch adjusts the PD for loans with poor
performance history, including loans that
are currently performing.

August 15, 2011
Structured Finance
before liquidation. This adjustment mainly influences the loans expected loss severity, since
many DQ loans already have an adjusted default probability at or near 100% in stressed
scenarios.
Sample Model Output
Fitchs model responds to varying loan attributes in terms of PD, sustainable market value
decline, loss severity, and expected losses, which is demonstrated by three hypothetical pools
rated at two different points in an economic cycle, as shown in the table below.
The output shows how the model distinguishes pools with different credit profiles in different
environments. While the model output plays a significant role in the rating process, other
important considerations are factored into a transactions rating. Thus, the loss expectations
indicated below may be adjusted during the rating process to account for operational or
structural considerations. The three examples of prime mortgage pools consist of high credit,
average credit, and layered risk quality pools having characteristics as described below.

- High Credit Pool: The high credit mortgage pool reflects a credit profile similar to prime
loans currently being originated by major lenders. The credit characteristics are
distinguished by borrowers with credit scores above 720, CLTVs below 80%, and 100%
full documentation.
- Average Credit Pool: The average credit mortgage pool generally has credit
characteristics similar to the high credit pool but with lower credit scores and slightly
increased concentrations of more risky credit attributes.
- Layered Risk Pool: The layered risk mortgage pool contains borrowers with credit scores
similar to that of the average credit pool but contains more high-risk combinations, such as
fewer full documentation loans, more cash-out refinances, and loans with piggy-back
second liens.
For the sample model output, the three mortgage pools were analyzed using the model
assuming two different economic environments.
- Current Environment: The model output for the current environment is indicative of an
economic environment where a large part of the housing correction is complete and
macroeconomic indicators are stabilizing. The weighted average sMVDs are only slightly
higher than the stressed market value declines applied at each rating category, indicating
that prices are close to sustainable levels while the base case ERF is higher than the
floors, indicating some volatility in macroeconomic conditions.
- Overvalued Environment: The output for the overvalued environment assumes the loans
were originated in an environment when home values are inflated by 25%. The sMVDs
and ERFs are materially higher than the sMVD rating stresses and ERF floors, reflecting
increased risk of a home price correction and an economic downturn.
The weighted average loss amounts for the three pools are markedly different, reflecting the
models ability to distinguish the low risk characteristics of the high credit pool from those of the
average and layered risk quality pools. This is most evidenced by comparing the weighted
average PD of the three pools in the current environment scenario. The differences in the PD
are attributable to the underlying risk attributes, since the sMVD and LS are more reflective of
the market dynamics (housing price assumptions), which are held constant in this example.
The high credit quality pool has an 11.6% PD in the AAAsf scenario, roughly 28% and 54%
lower than that of the average and layered risk pools.
The model can distinguish between
high-quality pools with low-risk attributes
and those consisting of riskier credits.

August 15, 2011
Structured Finance
The macroeconomic conditions also play a role in the models output, as can be seen in the
differences in the loss percentages when comparing the same pool in different environments.
All three portfolios have higher sLTV values and PDs in the overvalued environment than do
their current environment counterparts, reflecting greater market risk and resulting in
significantly higher expected losses at every rating stress category. For example, the expected
loss in the AAAsf scenario for the high credit pool would increase to 12.6% in an overvalued
environment from 7.1% in the current home price environment.
While the expected losses increase notably as market risk increases, it is also important to
note how the relationship changes between the Bsf and the AAAsf rating stresses,
depending on the economic environment. For the high credit pool, the multiple from the Bsf
loss to the AAAsf is approximately 16.0x times in the current economic environment but
compresses to roughly 8.0x in a highly overvalued environment. This relationship is maintained
for all three pools, reflecting the frameworks more definitive AAAsf scenario, which is
expected to result in greater rating stability though various economic cycles.
The model also recognizes shifts in
economic cycles as loss protection
levels rise when indicators are signalling
an overheated economy and home
prices are rising above sustainable
levels.
Current Environment Overvalued Environment

Weighted
Average
sMVD (%)
Weighted
Average
Adjusted
CLTV (%)
Weighted
Average
ERF
Weighted
Average PD
(%)
Weighted
Average
LS (%)
Weighted
Average
Loss (%)
Weighted
Average
sMVD (%)
Weighted
Average
Adjusted
CLTV (%)
Weighted
Average
ERF
Weighted
Average PD
(%)
Weighted
Average
LS (%)
Weighted
Average
Loss (%)
High Credit Quality Pool High Credit Quality Pool
AAAsf 42.6 123.9 3.5 11.6 61.3 7.1 AAAsf 51.3 144.1 3.5 18.6 67.6 12.6
AAsf 38.1 115.1 3.0 8.1 53.1 4.3 AAsf 47.5 133.8 3.0 13.4 59.6 8.0
Asf 33.7 107.4 2.5 5.3 45.0 2.4 Asf 43.8 124.9 2.5 9.0 51.5 4.6
BBBsf 29.3 100.7 2.0 3.3 39.5 1.3 BBBsf 40.0 117.1 2.5 7.0 46.1 3.2
BBsf 24.9 94.8 1.6 2.1 34.5 0.7 BBsf 36.3 110.2 2.5 5.5 40.8 2.3
Bsf 20.5 89.5 1.4 1.5 29.7 0.4 Bsf 32.5 104.1 2.5 4.4 35.5 1.6

Average Credit Quality Pool Average Credit Quality Pool
AAAsf 42.6 123.9 3.5 16.3 60.6 9.9 AAAsf 51.3 144.1 3.5 25.3 67.2 17.0
AAsf 38.1 115.1 3.0 11.7 52.5 6.1 AAsf 47.5 133.8 3.0 18.8 59.2 11.1
Asf 33.7 107.4 2.5 7.9 44.4 3.5 Asf 43.8 124.9 2.5 13.1 51.1 6.7
BBBsf 29.3 100.7 2.0 5.0 38.9 2.0 BBBsf 40.0 117.1 2.5 10.4 45.8 4.7
BBsf 24.9 94.8 1.6 3.2 34.0 1.1 BBsf 36.3 110.2 2.5 8.2 40.4 3.3
Bsf 20.5 89.5 1.4 2.2 29.2 0.7 Bsf 32.5 104.1 2.5 6.7 35.2 2.3

Layered Risk Pool Layered Risk Pool
AAAsf 42.4 128.2 3.5 25.0 59.7 14.9 AAAsf 51.3 150.2 3.5 36.7 66.6 24.5
AAsf 37.9 119.1 3.0 18.6 51.6 9.6 AAsf 47.5 139.5 3.0 28.8 58.7 16.9
Asf 33.5 111.1 2.5 13.0 43.7 5.7 Asf 43.8 130.2 2.5 21.1 50.8 10.7
BBBsf 29.1 104.2 2.0 8.5 38.4 3.3 BBBsf 40.0 122.0 2.5 17.2 45.5 7.8
BBsf 24.6 98.0 1.6 5.5 33.7 1.9 BBsf 36.3 114.8 2.5 14.0 40.3 5.6
Bsf 20.2 92.6 1.4 4.0 29.1 1.2 Bsf 32.5 108.5 2.5 11.4 35.1 4.0
Credit Attribute Summary
Credit Attribute High Credit Average Credit Layered Risk
Original Loan Amount ($) 670,203 670,203 670,203
Property Value ($) 1,023,158 1,023,158 1,023,158
Fixed Rate (%) 100 100 100
Weighted Average FICO Score 770 745 745
FICO Score <720 (%) 0 20 20
FICO Score <700 (%) 0 3 3
Combined Loan to Value (%) 70 70 73
Loans with Piggyback Second (%) 11 11 29
Owner Occupied (%) 100 97 97
Cash-Out Refinance (%) 0 25 58
Single-Family Dwelling (%) 92 92 92
Full Documentation (%) 100 100 60
California Concentration (%) 27 27 38

August 15, 2011
Structured Finance
Sensitivity Analysis
Fitchs sensitivity analysis provides three levels of rating sensitivities to demonstrate how the
ratings would react to steeper market value declines than that assumed at issuance. The
various rating sensitivities include defined stresses and defined sensitivities. The implied rating
sensitivities are only indicative of some of the potential outcomes and do not consider other risk
factors to which the transaction is exposed or are considered during the surveillance process.
Defined Stresses
Defined stresses show the impact of
three defined stress assumptions where
the base sustainable home price level is
10, 20, and 30 percentage points lower
than that derived at deal issuance.
These assumptions result in higher
sLTVs and steeper sMVDs - the most
significant drivers of PD and LS in
Fitchs loss model. The table at right
shows the impact on ratings for each
additional defined stress to sustainable price declines for a hypothetical deal.
Defined Sensitivities
Defined sensitivities describe the stresses to the assumptions required to reduce a rating by
one full category, to non-investment grade, and to CCCsf. The variable being stressed in this
analysis is Fitchs sustainable home
price assumption. The percentage
points shown in the table to the right
reflect the additional market value
declines that would have to occur to
impact ratings for each defined
sensitivity for a hypothetical deal.
Fitch performs sensitivity analyses on
the ratings assigned to RMBS
transactions. Fitch will provide details
of these sensitivities for each transaction, which may include the following considerations.
- Rating sensitivity to increased loan-level sustainable market value decline assumptions.
- Ratings sensitivity to extended foreclosure and liquidation timeline assumptions.
- Ratings sensitivity to defaults of the largest loans in the pool with increased sustainable
market value decline assumptions.
- Rating sensitivity to increased market value declines for loans concentrated in a
geographic region.
Defined Sensitivities
(%)
Additional Decline in Sustainable Price Level
Original
Rating
One Full
Category
Non-Investment
Grade To CCCsf
AAAsf 11 37 66
AAsf 11 30 49
Asf 11 21 41
BBBsf 11 12 34
BBsf 12 25
Bsf 13 13

Defined Stresses
(%)
Additional Decline in Sustainable Price Level
Original
Rating 10% 20% 30%
AAAsf AAAsf AAsf BBBsf
AAsf AAsf Asf BBsf
Asf Asf BBBsf BBsf
BBBsf BBBsf BBsf Bsf
BBsf BBsf Bsf < Bsf
Bsf Bsf < Bsf < Bsf

August 15, 2011
Structured Finance
Appendix A: Data Analysis and Model Development
Fitch conducted its analysis with CoreLogic/LoanPerformance (Corelogic/LP) data on
mortgages that satisfied the following conditions:
- Classified as prime loan.
- Fixed-rate loan.
- First lien loan.
- Originated between 1991 and 2007.
- Excluded loans with interest-only features.
- Excluded loans with negative amortization features.
- Excluded loans with loans with balloon payments.
- Excluded loans with original LTV ratio > 110%.
- Excluded loans with original CLTV ratio < 10%.
- Excluded loans where state = Puerto Rico, Virgin Islands, Guam, or missing.
Applying these conditions resulted in a dataset of 2.2 million loans.
Fitch also merged the following data elements into the CoreLogic/LP loan-level data:
- Fitchs sMVD was assigned to each loan based on its origination date and the state where
the property is located.
- The economic risk variable provided by a third party (UFA) was assigned to each loan based on
its origination date and the zip code of where the property is located or at the state level.
- For documentation type and back-end debt-to-income ratio, Fitch used its own loan-level
database as a supplement for loans where LP data was sparsely populated.
All loan attributes were taken as of the origination date of the loan. In determining whether to consider
loans as defaults or nondefaults, the performance of each loan was tracked through April 2011.
The dataset for the traditional prime was a subset of overall development sample with the
following additional filters applied.
- FICO scores > 720.
- Full documentation.
- Original combined LTV <80% if owner occupied.
- Original combined LTV <70% if non-owner occupied.
Applying these additional filters resulted in a traditional prime dataset of 510,000 loans.
Total and Traditional Pool Characteristics
Total Sample Traditional Sample
Vintage
No. of
Loans
Average
Loan Amount ($)
Average
FICO Score
Average
OCLTV (%)
Full
Doc (%)
No. of
Loans
Average Loan
Amount ($)
Average
FICO Score
Average OCLTV
(%) Full Doc (%)
1991 49,408 300,106 697 72 61 185 214,983 768 69 100
1992 123,937 290,444 726 69 63 711 230,991 769 65 100
1993 239,580 296,529 732 71 71 2,307 284,896 770 66 100
1994 87,976 266,483 713 72 66 1,024 247,567 770 66 100
1995 56,409 273,055 713 77 72 4,274 308,459 757 72 100
1996 65,012 283,197 711 76 76 5,976 311,683 757 73 100
1997 123,674 291,144 720 75 75 27,088 317,971 758 72 100
1998 269,585 327,714 724 73 73 74,837 350,058 758 70 100
1999 147,517 337,172 721 74 69 34,723 364,905 758 71 100
2000 91,951 354,395 724 76 71 26,427 392,254 759 72 100
2001 217,989 425,671 730 70 77 88,628 436,909 761 68 100
2002 208,846 458,181 736 66 74 90,277 464,952 763 64 100
2003 212,061 482,628 739 65 57 66,480 487,763 763 63 100
2004 100,606 475,548 738 69 65 30,728 499,316 763 66 100
2005 92,167 528,104 740 71 54 23,913 573,989 766 68 100
2006 81,339 567,079 741 75 50 18,432 632,497 768 70 100
2007 71,097 594,501 743 76 46 14,618 670,518 770 70 100

August 15, 2011
Structured Finance
Appendix B: Roll Rate Analysis Description
The metrics used to quantify the five-year observed performance were roll rates to a worse DQ
bucket, cure rates to a better DQ bucket, default rates, and prepayment rates. For each metric,
the monthly average was calculated over the observation period of April 2006 through
April 2011.
Fitch applied a granular approach to distinguish between loans with different attributes. Fitchs
methodology established 128 cohorts, each defined by a unique combination of loan attributes.
The attributes consisted of two documentation buckets, four mark-to-market combined loan-to-
value ratio (MTM CLTV) buckets, four FICO buckets, and four payment history buckets (2 x 4 x
4 x 4 = 128). Each loan fell into exactly one cohort, based on its attributes. For example, a loan
that had full documentation, a 90% MTM CLTV, a 700 FICO score, and a clean payment
history would belong to one cohort, while a loan with the same documentation type, CLTV, and
FICO score but with a D payment history would belong to a different cohort.
After defining the cohorts and populating each one with loans from its sample, Fitch then
calculated a unique set of performance histories for loans within each cohort. The result was
128 sets of two-year average roll, cure, default, and prepayment rates for each of the 128
cohorts. The next step was to arrive at cohort-specific starting points from which to start Fitchs
projections. Using April 2011 data, the agency took a snapshot of the current loan count,
delinquency distribution, average CLTV, and geographic distribution for each cohort. Lastly, the
average observed performance rates were applied to the starting points, projecting future
performance on a month-by-month basis.
While cohort-specific performance rates were applied at a constant rate, each cohorts MTM
CLTV designation was dynamic based on several projections. The loans were projected to
amortize based on their schedules, and an annual inflation rate of 1.75% was assumed, both of
which worked to improve the cohorts MTM CLTV designation over time. Additionally, Fitch
incorporated state-level housing price forecasts and arrived at an aggregate forecast number
for each cohort based on its state distribution. Insofar as these three forecasts shifted the
cohorts average CLTV beyond the definition of the cohort, the performance rates of a different
cohort one whose definition includes the new CLTV would take over. For example,
assume cohort A is defined in part by having a CLTV range of 80%100%. Snapshot A is the
starting point for the loans in cohort A and has a CLTV of 90%. Cohort As performance rates
would be applied to snapshot A monthly. However, further assume in month 30 that the home
price forecast raises the CLTV of snapshot A to 101% from 90%. Since it no longer conforms to
cohort As definition, cohort As performance would no longer apply. Snapshot A now falls into
cohort Bs definition of having a CLTV range of 100%140%, so starting in month 31, cohort
Bs performance rates would be applied to snapshot A.
These sometimes-shifting performance rates were projected monthly until all loans in the
starting snapshots either paid off or defaulted. This process was repeated for each of the 128
cohorts. To translate the cohort-level results into loan-level data for input into the regression,
loans within each cohort were randomly flagged as defaulters or nondefaulters, according to
the cohort-level projected default rate. For example, if a cohort consisted of 100 outstanding
loans, and the roll rate analysis projected a 25% default rate, 25 of the 100 loans would be
randomly assigned as defaulters and 75 as nondefaulters. The final result was a one or a zero
assigned to each outstanding loan to represent default or nondefault for the regression.

August 15, 2011
Structured Finance
Appendix C: Regression Data and Methodology
Loan-level data was extracted from the Loan Performance database. Of the 2.2 million loans
dataset described in Appendix A, a subset of 1.2 million loans was used to construct the PD
regression model after the sample was balanced for high issuer and nontraditional prime loan
concentrations in certain vintages. The balanced sample was further weighted by vintage to
prevent overrepresentation of high refinance volume vintages such as 1998, 2001, 2002, or 2003.
This further helped take into account changes in the default behavior of peak default vintages.
Logistic Regression: The logistic regression model was specified to estimate the drivers of
prime, fixed-rate loan PD. The default risk for mortgage loan is characterized as a dichotomous
event that can take one of two possible outcomes, i.e. default or nondefault. The logit model is
based on the cumulative logistic probability function where it can be specified as a generalized
linear function:

))] ; ( 1 log(( ) 1 ( ) ; ( log [ ) ; ( log | | |
i i
N
i i i
x p y x p y x L + =

Where g is the link function through which probability of default (p, equivalent to probability of
default) is related to the explanatory variable X, p is the probability of event, prob(Y=1) is the
parameter vector, and X is the matrix of explanatory variables. The expected probability of
default can be analytically solved for p through logit transformation where
Where is the logistic distribution function. (.) A
The logistic regression specified above assumes a natural logarithmic relationship between the
explanatory variables and the ratio of event (foreclosure) to non-event (no foreclosure),
otherwise known as the odds ratio. Thus the natural log of the odds is linearly related to the
explanatory variables.
This form is desirable in that the logit function, g(p), is linear in its parameters, and the logistic
transformation always generates an outcome p probability between 0 and 1. Estimates of the
parameters are obtained by maximizing the sample log likelihood function:
where L(.) is the likelihood function, yi=1 for a defaulted loan and 0 for a nondefaulted loan, and
p is the probability measure.
SAS and other statistical software contain statistical procedures to perform logistic regression.
The SAS procedure LOGISTIC is used to estimate the parameters of the regression model with
options to generate goodness of fit and other diagnostics designed to assess model robustness.
1 i=
( )
( ) |
|
|
X
X
X Y prob p
exp 1
exp
) ( ) 1 (
+
= A = = =
| X p p p g = = )) 1 /( ln( ) (

August 15, 2011
Structured Finance
Univariate analysis is initially performed for each variable individually to understand
distributions, strength, and forms of relationships to default before performing the analysis in a
multivariate context. Each variable is assessed for the need of additional transformations
based on the nature of relationship to default. Nonlinear transformations are applied when
necessary to improve the fit of the regression. Performance of the regression was measured by
nonparametric statistics for rank ordering/separation ability (i.e. Somers D, Kolmogorov-
Smirnov, or K-S). The tests are performed on all 2.2 million data and on the traditional loan
sample only in line with target portfolios. All tests showed the model had high power in rank
ordering and separation ability. The main focus in measuring the models performance is the
point-estimate accuracy on benchmarked vintage projected default rates where average rates
of predicted versus observed foreclosures were measured and compared for loans originated
from 1991 through 2007. The model successfully captured the dramatic transition in credit risk
for the peak default vintages of 2005, 2006, and 2007, as seen in the chart below for traditional
loan sample.

0
20
40
60
80
100
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
0
2
4
6
8
10
12
No. of Loans Tr adi t i onal PD Act ual and Pr oj ect ed New Model
Model Perf ormance: Act ual vs. Predi ct ed Tradi t i onal Sampl e
(000) (%)
0
50
100
150
200
250
300
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
0. 00
5. 00
10. 00
15. 00
20. 00
25. 00
No. of Loans Al l Sampl e PD Act ual and Pr oj ect ed New Model
Model Perf ormance: Act ual vs. Predi ct ed Al l Sampl e
(000) (%)

August 15, 2011
Structured Finance
Appendix D: Economic Risk Factors
The impact of economic factors on future defaults and losses is captured by the National Risk
Index (NRI) and regional risk multipliers provided by UFA, which are indicators of home price
forecasts and a number of other econometric measures. UFAs analysis is used to derive the
quarterly UFA Mortgage Report risk multipliers on a state, zip code, and national level. The
UFA multipliers are incorporated into Fitchs new PD model and raise or lower the expected
default probability on a given mortgage loan to reflect national and regional economic risk
forecasts.
State and zip code-level risk multipliers represent the level of expected risk over the life of a
loan relative to the national average on a constant quality basis. For example, if the UFA
default multiplier for a state is 0.90, expected defaults in that state are 90% of those for the
average loan in the U.S.
Supplementing the state-level risk multipliers with zip code multipliers provides increased
granularity in the default and loss risk analysis, since conditions at the regional level can exhibit
more disparate trends than those indicated at the state level due to variations in industry
concentration/diversification, employment growth, personal income, demographics, and other
factors. Both national and regional components are applied to the PD, aligning Fitchs base
case and stressed expected loss for newly originated loans with prevailing and forecast
economic conditions.
National Risk Index
The NRI provides a default forecast for loans originated today relative to loans underwritten
during the 1990s. Thus, the 19902000 economic climate serves as a benchmark from which
todays metrics are measured and is the comparative basis for future defaults. UFA assumes
the quality of the borrower, loan characteristics, and legal environment remain constant; thus,
changes in the index reflect only changes in current macroeconomic conditions.
Each quarter, UFA evaluates economic conditions in the U.S. and assesses how those
conditions will impact future defaults, prepayments, loss recoveries, and loan values for
nonprime loans. The NRI reflects changes in economic measures, such as real GDP growth,
real consumer spending, business spending, national unemployment rates, CPI inflation rates,
mortgage rates, national house price appreciation, and housing permits.
Because UFA updates the index quarterly, credit enhancement levels for pools with similar
characteristics could vary from quarter to quarter. To reflect the risk of further economic
deterioration and to provide for more stable credit enhancement and ratings through the
economic cycle, stressed index value floors are applied to all investment-grade rating
categories. The stressed level will be evaluated with each quarterly update from UFA. Should
the NRI rise more than expected, the stressed multiplier may be increased accordingly. If and
when the NRI begins to decrease, Fitch will re-evaluate the stress and may adjust it
accordingly. Any changes to the stressed level will be announced in the quarterly Fitch
research, which will be available at www.fitchratings.com.
Regional Risk Multipliers
UFAs analysis of regional risk takes into account state and local economic metrics, such as
personal income and distribution, employment growth, housing construction, and other

August 15, 2011
Structured Finance
indicators. It also factors in a demographic component, which includes unemployment rates
and population growth, as well as a political component that considers local taxes and zoning
regulations. The resulting multipliers reflect the areas expected lifetime default risk in
comparison to the national average. For example, a multiplier of 1.10 indicates that loans
originated in this region have a 10% higher default risk than the national average in the 1990s.
The regional risk multipliers are updated quarterly and are available at www.fitchratings.com.
Variability in credit enhancement levels could occur if there are significant changes in regional
concentrations among pools with similar credit characteristics. Also, unlike the NRI, where one
value is applied to all loans, risk multipliers vary by location; thus, regions with rising risk
multipliers could offset those declining in risk.

0. 0
0. 5
1. 0
1. 5
2. 0
2. 5
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010
Hi st ori cal Nat i onal Ri sk Index
Sour ce: Uni ver si t y Fi nanci al Associ at es, LLC .

August 15, 2011
Structured Finance
Appendix E: Heat Maps for Sustainable Home Price Model (2006 and Current)
Overvaluation in 2011 as Predicted by Sustainable Home Price Model
At or near sustainable
10-20%
20-30%
30-40%
40%+
Overvaluation in 2006 as Predicted by Sustainable Home Price Model
At or near sustainable
10-20%
20-30%
30-40%
40%+

August 15, 2011
Structured Finance
Appendix F: Summary of Changes to Model Exposure Draft

Model Component Exposure Draft Final Version
Regression Analysis
1. Roll Rate Methodology
Default projections for recent vintage collateral were based on
delinquency roll rates, defaults and prepayment performance over the
past two years to extrapolate future defaults. Fitch assumed the behavior
would continue at the same rate for the remainder of the life of the pools.
Fitch's default projections were updated to consider actual performance through
April 2011. The agency's projections for recent 'peak' vintages are based on five
year prior experience to capture some pre-crisis performance, reflecting Fitch
expectation of future stabilization in default performance.
Frequency Of Foreclosure Variables
1. Sustainable Loan to Value Ratio (SLTV)
SLTV was not an explicit PD variable in the agency's exposure draft.
The agency's view on current sustainable borrower equity was caputured
by two separate PD variables - Sustainable Market Value Decline and
Original Combined Loan-to-Value Ratio.
The Sustainable Loan to Value Ratio (SLTV) was included in the agency's final PD
regression model. This variable effectively combines sMVD and OCLTV to provide
a view on borrower sustainable equity, which is the most explanatory variable in the
agency's new PD regression model.
2. Cash-out Refinances 68% PD penalty relative to purchase loan baseline.
PD penalty increased to 100% relative to purchase loan baseline based on new
regression results.
3. Rate/Term Refinances 26% PD penalty relative to purchase loan baseline.
PD penalty increased to 30% relative to purchase loan baseline based on new
regression results.
4. Loan Term
15 year loans applied a 40% PD reduction relative to 30 year term
baseline: Loan terms over 30 years were assigned a 100% PD penalty
relative to the 30-year baseline.
15 year loans applied a 50% PD reduction relative to 30 year term baseline: Loan
terms over 30 years were assigned a 150% PD penalty relative to the 30-year
baseline.
5. Investor (Non-Owner-Occupied Properties) 20% PD penalty relative to owner-occupied baseline.
35% PD penalty relative to owner-occupied baseline based on new regression
results.
6. Risk Premium
Risk premium was the eighth ranked PD variable in the regression
model. This variable sought to assess credit risk by comparing the
mortgage note rate relative to prevailing mortgage at loan orgination.
Risk premium was removed from the final PD regression model due to due to multi-
colinearity with other regression variables as well as concerns about its
predictiveness in the current environment. For example, at present, loan mortgage
rates may reflect non-credit issues including mortgage liquidity constraints in
specific segments of the market.
7. Seasoning Treatment for seasoned loans was not included in the exposure draft.
When estimating losses on seasoned mortgage loans, Fitch uses the same PD, LS
and rating stress framework as that used for analyzing newly-originated loans.
However, two key adjustments based on the borrowers updated sLTV and
payment history are made to account for the additional observed performance data
that is available for seasoned loans. Depending on Fitchs view of the borrowers
equity position and the loans performance over time, these adjustments can either
increase or decrease the loans loss expectation compared to the loss level the
model would have been assigned at loan origination.
Loss Severity
1. Quick Sale Adjustments
Quick sale adjustment percentages varied based on rating stress
ranging from 20% in 'AAA' and 10% in 'B'.
Extensive research was conducted into distressed sale discounts over time, and
there was no evidence to suggest that properties sell at a greater discount in higher
stress scenarios. While in higher stresses, more properties will be affected by the
quick-sale, individual property adjustments have remained relatively consistent on
liquidating loans historically at approximately 15% below market value. This haircut
is now applied to all rating categories.
2. Foreclosure Costs - Taxes and Insurance
The agency estimated tax and insurance costs at 2% of the appraised
value, and modelled these as a one-time cost.
The updated approach takes a dynamic view of taxes and insurance, which reflect
costs that will be incurred over the entire liquidation period, or up to 3 years in the
AAA stress. On average, these costs are approximately 2% of the property value
per year based, but are variable by state. Maintenance costs are now assumed to
be a one-time cost equal to 1% of the property value, plus 0.25% per year over the
liquidation period.
Rating Stresses
1. Sustainable Market Value Decline (SMVD) Floors
Assumed 45% at 'AAA' through 20% at 'B' by benchmarking Fitch's peak-
to-trough 35% market value decline expectation for recent correction to a
'A' stress.
The agency now applies a dynamic two-step process whereby properties are first
adjusted to their sustainable values and then applied further sMVD stresses that
correspond to different rating scenarios. This approach better considers where we
are in the housing cycle, with stresses and CE increasing in 'unsustainable bubble
scenarios' and declining in environments where prices are approaching
sustainability.
2. Liquidation Timelines
29 month foreclosure and liquidation timelines for the 'B' through 'A"
rating categories were established based on observations in current
environment and assumed further extended timelines for 'peak' vintages.
'AAA' and 'AA' timelines incorporated additional timelines extensions of 6
and 12 months, respectively.
Foreclosure and liquidation timelines were shortened in the lower rating categories
to reflect Fitch's expectation that timelines will decrease once distressed inventory
is liquidated and servicing procedures and controls are fully implemented. High
investment grade stresses were only reduced slightly.
3. Economic Risk Factor (ERF) Floors
ERF floors based on UFA's 2010 national risk index values. ERF floors
for all investment grade categories were set at the peak level of 4.7
representing 2007 stressed environment.
ERF floors are based on UFAs 2Q2011 National Risk Index Values, which includes
a revised scale. The agency decided to tier the ERF floor stresses throughout the
capital structure to increase stability in high investment grade categories. While the
ERF floor of 2.5 at the 'A' category continues to represent the 2007 economic
environment (equivalent to 4.7 in old UFA scale), 'AAA' through 'AA' floors
represent more stressful economic conditions.

Structured Finance

ALL FITCH CREDIT RATINGS ARE SUBJECT TO CERTAIN LIMITATIONS AND DISCLAIMERS. PLEASE READ THESE
LIMITATIONS AND DISCLAIMERS BY FOLLOWING THIS LINK:
HTTP://FITCHRATINGS.COM/UNDERSTANDINGCREDITRATINGS. IN ADDITION, RATING DEFINITIONS AND THE
TERMS OF USE OF SUCH RATINGS ARE AVAILABLE ON THE AGENCYS PUBLIC WEB SITE AT
WWW.FITCHRATINGS.COM. PUBLISHED RATINGS, CRITERIA, AND METHODOLOGIES ARE AVAILABLE FROM
THIS SITE AT ALL TIMES. FITCHS CODE OF CONDUCT, CONFIDENTIALITY, CONFLICTS OF INTEREST, AFFILIATE
FIREWALL, COMPLIANCE, AND OTHER RELEVANT POLICIES AND PROCEDURES ARE ALSO AVAILABLE FROM
THE CODE OF CONDUCT SECTION OF THIS SITE.
Copyright 2011 by Fitch, Inc., Fitch Ratings Ltd. and its subsidiaries. One State Street Plaza, NY, NY 10004.Telephone:
1-800-753-4824, (212) 908-0500. Fax: (212) 480-4435. Reproduction or retransmission in whole or in part is prohibited except
by permission. All rights reserved. In issuing and maintaining its ratings, Fitch relies on factual information it receives from
issuers and underwriters and from other sources Fitch believes to be credible. Fitch conducts a reasonable investigation of the
factual information relied upon by it in accordance with its ratings methodology, and obtains reasonable verification of that
information from independent sources, to the extent such sources are available for a given security or in a given jurisdiction.
The manner of Fitchs factual investigation and the scope of the third-party verification it obtains will vary depending on the
nature of the rated security and its issuer, the requirements and practices in the jurisdiction in which the rated security is offered
and sold and/or the issuer is located, the availability and nature of relevant public information, access to the management of the
issuer and its advisers, the availability of pre-existing third-party verifications such as audit reports, agreed-upon procedures
letters, appraisals, actuarial reports, engineering reports, legal opinions and other reports provided by third parties, the
availability of independent and competent third-party verification sources with respect to the particular security or in the
particular jurisdiction of the issuer, and a variety of other factors. Users of Fitchs ratings should understand that neither an
enhanced factual investigation nor any third-party verification can ensure that all of the information Fitch relies on in connection
with a rating will be accurate and complete. Ultimately, the issuer and its advisers are responsible for the accuracy of the
information they provide to Fitch and to the market in offering documents and other reports. In issuing its ratings Fitch must rely
on the work of experts, including independent auditors with respect to financial statements and attorneys with respect to legal
and tax matters. Further, ratings are inherently forward-looking and embody assumptions and predictions about future events
that by their nature cannot be verified as facts. As a result, despite any verification of current facts, ratings can be affected by
future events or conditions that were not anticipated at the time a rating was issued or affirmed.
The information in this report is provided as is without any representation or warranty of any kind. A Fitch rating is an opinion
as to the creditworthiness of a security. This opinion is based on established criteria and methodologies that Fitch is
continuously evaluating and updating. Therefore, ratings are the collective work product of Fitch and no individual, or group of
individuals, is solely responsible for a rating. The rating does not address the risk of loss due to risks other than credit risk,
unless such risk is specifically mentioned. Fitch is not engaged in the offer or sale of any security. All Fitch reports have shared
authorship. Individuals identified in a Fitch report were involved in, but are not solely responsible for, the opinions stated therein.
The individuals are named for contact purposes only. A report providing a Fitch rating is neither a prospectus nor a substitute for
the information assembled, verified and presented to investors by the issuer and its agents in connection with the sale of the
securities. Ratings may be changed or withdrawn at anytime for any reason in the sole discretion of Fitch. Fitch does not
provide investment advice of any sort. Ratings are not a recommendation to buy, sell, or hold any security. Ratings do not
comment on the adequacy of market price, the suitability of any security for a particular investor, or the tax-exempt nature or
taxability of payments made in respect to any security. Fitch receives fees from issuers, insurers, guarantors, other obligors,
and underwriters for rating securities. Such fees generally vary from US$1,000 to US$750,000 (or the applicable currency
equivalent) per issue. In certain cases, Fitch will rate all or a number of issues issued by a particular issuer, or insured or
guaranteed by a particular insurer or guarantor, for a single annual fee. Such fees are expected to vary from US$10,000 to
US$1,500,000 (or the applicable currency equivalent). The assignment, publication, or dissemination of a rating by Fitch shall
not constitute a consent by Fitch to use its name as an expert in connection with any registration statement filed under the
United States securities laws, the Financial Services and Markets Act of 2000 of Great Britain, or the securities laws of any
particular jurisdiction. Due to the relative efficiency of electronic publishing and distribution, Fitch research may be available to
electronic subscribers up to three days earlier than to print subscribers.
August 15, 2011

FI US Re Sip Rime Loan Loss Model Criteria 110815

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

FI US Re Sip Rime Loan Loss Model Criteria 110815

Uploaded by

Copyright:

Available Formats

Structured Finance

Residential Mortgage / U.S.A.

You might also like